JPH0797835B2 - Ghost removal device - Google Patents

Description

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a ghost elimination device for a television received signal using a transversal filter, and more particularly, to shorten the time required for correlation calculation for detecting a ghost. The present invention relates to such a ghost removing device.

[Conventional technology]

In a television receiver, when a radio wave (desired wave) directly coming from a transmitting antenna and a radio wave reflected from a building are received by the receiving antenna at the same time, the image due to the desired wave and the image due to the reflected wave are deviated. The so-called ghost appears.

Such a ghost component for a television receiver is a major factor in deteriorating the image quality, and various measures have conventionally been taken to remove or prevent the ghost component. As one of them, a ghost removing device using a transversal filter in a video band is known.

This device connects a large number of delay elements each having a minute delay time determined by the highest frequency component contained in the video signal in series, weights the output of each delay element by a coefficient circuit, then adds and outputs, A ghost-removed video signal (ghost compensation signal) is obtained, and the above-described configuration immediately forms a transversal filter configuration.

Although there are various algorithms for determining the coefficient (hereinafter, referred to as tap gain) of the above-described coefficient circuit which constitutes such a transversal filter, from the viewpoint of stability, eyeEeTransactionsOnConsumer・ Electronics volume CE-26, 1980 8
Mon, pages 629 to 635 (IEEE, Transactions on Consu
mer Electronics, Vol.CE-26, August 1980, P.629 ~ P.63
The correlation method algorithm discussed in 5) is generally used (a concrete example of the ghost removing device is described in JP-A-54-70725).

In this correlation algorithm, the tap gain Ci (i means the i-th coefficient circuit) is determined by the following equation.

Where α is a positive constant {X _K }: input waveform to the transversal filter {Y _K }: output waveform from the transversal filter That is, a new tap gain (corrected tap gain) Ci, in the i-th coefficient circuit Ci, new is obtained from the correlation value between the input waveform {X _K } and the output waveform {Y _K }.

FIG. 2 shows a block diagram of a conventional ghost removing device which determines the tap gain of the coefficient circuit of the transversal filter by the above-mentioned correlation method algorithm.

In the figure, 1 is a video signal input terminal, 2 is a video signal output terminal, 3 is a transversal filter, 4 is a sync separation circuit that separates and outputs a sync signal from the input video signal, and 5 is a reference for the sync signal. , A timing generation circuit for generating a timing signal, 6 and 7 are A / D conversion circuits, 18 is a subtraction circuit, 19 is a tap gain memory, 20 is a D / A conversion circuit, 21 and 22 are differential circuits, and 23 and 24 and 27 are switch circuits, 25 is an input waveform memory, 26 and 28 are output waveform memories, 29 is a multiplier, and 30 is a cumulative adder.

The video signal input from the video signal input terminal 1 is output from the video signal output terminal 2 after the ghost component is removed (reduced) by the transversal filter 3.

The input video signal and the output signal of the transversal filter 3 are A / D converted by the A / D conversion circuits 6 and 7 via the differentiation circuits 21 and 22, respectively, and then input to the switch circuits 23, 24 and 27, and input. The waveform memory 25 and the output waveform memories 26 and 28 are input to a correlator M including a multiplication circuit 29 and a cumulative adder 30.

In the correlator M, the second term on the right side of the equation (1) is obtained, and the output signal thereof is subtracted from the tap gain (Ci, old) stored in the tap gain memory 19 by the subtraction circuit 18,
Get a new tap gain (Ci, new). This tap gain is stored in the tap gain memory 19 and, at the same time, D / A converted by the D / A conversion circuit 20 and then input to the transversal filter 3 and given to the i-th coefficient circuit therein.

Hereinafter, the operation of the correlator M will be described in detail. In order to remove the ghost, it is necessary to detect the position where the ghost component exists, but since the leading edge of the vertical synchronizing signal included in the video signal is widely used as the reference position signal for that purpose, This case will be described with reference to FIG.

FIG. 3 (a) shows a waveform diagram in the vicinity of the leading edge E of the vertical synchronizing signal V.SYNC when a single in-phase ghost G is added.

By closing the switch circuits 23, 24, 27 to the side a, the differential waveforms of the input / output signals of the transversal filter 3 are taken into the input waveform memory 25 and the output waveform memories 26, 28, respectively. At this time, the input waveform memory 25 is shown in FIG.
Storing data of the m samples before and after X ₁ through XM of the leading edge E of the vertical synchronizing signal as indicated by.

Note that this sample data is A / D based on the timing signal (clock signal) output from the timing generation circuit 5.
It goes without saying that the D conversion circuit 6 operates to result in sampled data.

As shown in FIG. 3 (c), the output waveform memory 28 stores m sample data (Y) at the same relative position as the sample data stored in the input waveform memory 25 among the output signals of the transversal filter 3. _{1 to} Ym) is stored.

Further, the output waveform memory 26 includes (N−m) sample data (Ym) after the sample data stored in the output waveform memory 28 among the output signals of the transversal filter 3.
_{+1 to} Y _N ) is stored. Of course, this sample data is data sampled by the A / D conversion circuit 7.

After capturing the sample data in the input waveform memory 25 and the output waveform memories 26 and 28, the switch circuits 23 and 27 are closed to the side b, and the sample data stored in the input waveform memory 25 and the output waveform memory 28 are sequentially read, Input to the multiplication circuit 29 and the accumulation circuit 30. The output of the accumulator 30 Is obtained.

At this time, since the switch circuits 23 and 27 are closed to the b side,
At the end of the calculation of the equation (2), the data at the start of the calculation are taken into the input waveform memory 25 and the output waveform memory 28.

Next, the switch circuit 24 is closed to the b side, the switch circuit 27 is closed to the a side, the data in the output waveform memories 26 and 28 are read one by one, and Y _{2 to} Ym _{+1 in} the output waveform memory 28 and the output waveform memory 26 in the output waveform memory 26. Y
Data of m ₊₂ to Y _N and Y ₁ are loaded respectively. After that, the switch circuit 23 is closed to the b side, the switch circuit 27 is closed to the b side, and the respective data of the input waveform memory 25 and the output waveform memory 28 are sequentially taken out, so that the output of the cumulative adder 30 is obtained. To get

After that, by repeating the same operation, the cumulative adder 30
The second term on the right side of the equation (1) is obtained from the output of Therefore, by subtracting the data of the tap gain memory 19 and the data of the cumulative adder 30 in the subtractor 18, the operation of the above formula (1) is executed.

[Problems to be solved by the invention]

In the above-mentioned conventional technique, since the correlation calculation represented by the equation (1) is executed on the time axis, there is a problem that the calculation time becomes very long.

An object of the present invention is to significantly reduce the calculation time in the ghost elimination device and shorten the convergence time required for ghost elimination.

[Means for solving problems]

The above-mentioned object is achieved by performing the correlation calculation shown in the above equation (1) not on the time axis but on the frequency axis using the fast Fourier transform (FFT).

[Action]

Generally, the cross-correlation function of the two signals x (t) and y (t)
Rxy (t) is expressed by the following equation.

Rxy (t) = ▲ ^∞− _∞ ▼ x (t) ・ y (t−τ) dt
(4) Here, assuming that the Fourier transforms of x (t), y (t) and Rxy (t) are X (ω), Y (ω) and Rxy (ω), the following equation holds.

Rxy (ω) = X (ω) · Y ^* (ω) (5) where Y ^* (ω) is a conjugate complex number of Y (ω) Therefore, the Fourier transform of the input / output signals of the transversal filter is performed by the fast Fourier transform. The above equation (1) can be realized by using (FFT).

〔Example〕

 An embodiment of the present invention will be described below with reference to FIG.

FIG. 1 is a block diagram showing a correlator M as a main part in one embodiment of the present invention.

In the figure, 8 is an input waveform memory, 9 is an output waveform memory, 10 and 11 are Fourier transform circuits respectively, 12 is a conjugation circuit for obtaining a conjugate complex number, 16 is a multiplication circuit, and 17 is an inverse Fourier transform circuit.

Sample data in the vicinity of the leading edge of the vertical synchronizing signal of the input and output signals of the transversal filter 3 is loaded into the input waveform memory 8 and the output waveform memory 9. Fourier transform circuit
The Fourier transform X (ω) of the input signal is obtained by 10 and the Fourier transform Y (ω) of the output signal is obtained by the Fourier transform circuit 11.
Is obtained, and the conjugation is performed by the conjugation circuit 12 to obtain the Fourier transform Y ^* (ω). In the multiplication circuit 16, the outputs of the Fourier transform circuit 10 and the conjugation circuit 12 are multiplied to obtain X
After obtaining (ω) · Y ^* (ω), it is converted into data on the time axis by the inverse Fourier transform circuit 17 to obtain the correlation between the input signal and the output signal.

Below, the output of the inverse Fourier transform circuit 17 is converted to the subtraction circuit of FIG.
By inputting to 18, each tap gain Ci of the transversal filter is obtained.

FIG. 4 is a block diagram showing another configuration example of the correlator M as an essential part of one embodiment of the present invention.

In the figure, blocks with the same numbers as in FIG. 1 indicate the same functional blocks. 31 is a CPU and 32 is a RAM.

In the correlator M shown in FIG. 4, the sample data taken in the input waveform memory 8 and the output waveform memory 9 is input to the CPU 31, and the CPU 31 performs Fourier transform, multiplication and inverse Fourier transform by software, and the obtained result is obtained. Is stored in the RAM 32 and output to the subtraction circuit 18 in FIG.
A program for operating the CPU 31 is stored in the RAM 32.

In the correlator M shown in FIG. 4, various operations are executed by software, so that the correlator M shown in FIG.

In the correlator M shown in FIGS. 1 and 4, mutual calculation of the input signal and the output signal can be executed on the frequency axis by using the Fourier transform.

Therefore, if the Fast Fourier Transform (FFT) is used, the operation time can generally be shortened by N / log ₂ N (N is the number of data), so that the tap gain Ci of the transversal filter 3 can be obtained at high speed. Become.

In the correlator M shown in FIG. 1, the correlation between the input signal x (t) and the output signal y (t) is obtained by the above equation (4), but it may be obtained by the following equation (6).

Rxy ′ (t) = ▲ ^∞− _∞ ▼ x (t−τ) ・ y (t) dt
(6) In this case, the Fourier transform Rxy ′ of the above correlation Rxy ′ (t)
Since (ω) is Rxy ′ (ω) = X ^* (ω) · Y (ω) (7), the conjugation circuit 12 in FIG. 1 may be configured on the Fourier transform circuit 10 side.

When the input signal x (t) or the output signal y (t) is expressed as an even function, Therefore, the conjugate circuit 12 can be omitted.

Next, the correspondence between the correlation calculation result and the tap coefficient of the transversal filter 3 will be described with reference to FIGS. 5 and 6.

FIG. 5 is a block diagram showing a configuration example of the transversal filter 3.

In the figure, 50 is a delay line with taps having a tap interval τ (delay time τ between taps), 51 is a coefficient circuit, and 52 is an addition circuit.

A video signal input from the video input terminal 1 is delayed by a delay line with a tap 50 for a predetermined time, and then gain controlled by a coefficient circuit 51, and then added by an adder circuit 52 and output. .

FIG. 6 shows a reference position signal for ghost removal,
FIG. 7 is a waveform diagram showing how to obtain each coefficient Ci of the coefficient circuit 51 when the leading edge E of the vertical synchronizing signal which is generally widely used is used.

It is assumed that an in-phase ghost component G having a delay time mτ and an amplitude A from the reference signal as shown in FIG. 6A is added to the input video signal.

Set the sampling interval of the A / D conversion circuits 6 and 7 (see Fig. 2) to τ
Then, when 1 pieces of data (X _{1 to} Xl) before and after the leading edge E of the vertical synchronizing signal are loaded into the input waveform memory 8 (see FIGS. 1 and 4), the data are shown in FIG. 6 (b). ) Is represented by Here, the impulse I is the differential output of the leading edge E of the vertical synchronizing signal. Similarly, the output waveform memory 9, FIG. 6 N data as shown in _{(c) (Y 1 ~Y N} )
Is captured.

From the waveforms of FIGS. 6 (b) and 6 (c), the correlation calculation shown in the above equation (5) is performed on the frequency axis, and the result is returned to the time axis using the inverse Fourier transform. The result shown in (d) is obtained.

That is, A ₀ , which is the correlation between the differential waveforms I of the leading edge E of the vertical synchronizing signal, and Am, which is the correlation between the waveform of FIG. 6 (b) and the ghost differential waveform G ′ shown in FIG. 6 (c). In all other cases, the correlation value is 0 except that That is, the location of the ghost component when the leading edge E of the vertical synchronizing signal is used as the reference position has been found.

Therefore, by obtaining each coefficient Ci of the coefficient circuit 51 of the transversal filter 3 in FIG. 5 by Ci, new = Ci, old−Ai (6), the coefficient circuit of the transversal filter 3 is removed so as to remove the ghost component. 51 can be controlled.

〔The invention's effect〕

As described above, according to the present invention, the required tap gain Ci of the coefficient circuit that constitutes the transversal filter can be obtained at high speed, so that the convergence time required for ghost removal can be greatly shortened. A removal device can be provided.

[Brief description of drawings]

FIG. 1 is a block diagram showing a correlator as an essential part of one embodiment of the present invention, FIG. 2 is a block diagram showing a conventional example of a ghost removing device, and FIG. 3 is an operation of the correlator shown in FIG. 4 is a block diagram showing another example of the configuration of a correlator as an essential part of one embodiment of the present invention, FIG. 5 is a block diagram showing an example of the configuration of a transversal filter, and FIG. The figure is a waveform diagram for explaining the relationship between the coefficient of the coefficient circuit and the correlation value. Explanation of code 3 ... Transversal filter, 8 ... Input waveform memory, 9 ... Output waveform memory, 10,11 ... Fourier transform circuit, 12 ... Conjugation circuit, 16 ... Multiplication circuit, 17 ... Inverse Fourier transform circuit, 31 …… CPU, 32 …… RAM

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Hiroshi Shibuya, 292 Yoshida-cho, Totsuka-ku, Yokohama, Kanagawa Nikko Video Engineering Co., Ltd. (56) References JP 59-198087 (JP, A) JP JP 60-91773 (JP, A) JP-A-61-198978 (JP, A)

Claims (2)

[Claims] 1. A transversal filter which is supplied with an input video signal and reduces and outputs a ghost component therein, and a signal including a reference position signal which serves as a reference for detecting the position of the ghost component in the video signal. From the input signal of the transversal filter to store the sample data, and a signal including the reference position signal and the ghost component from the output signal of the transversal filter to store the sample data. And a correlator that takes out the sample data stored in the first storage circuit and the sample data stored in the second storage circuit and calculates the cross-correlation function between them. The tap gain of the coefficient circuit that constitutes the transversal filter is output to the correlator. And a tap gain correction circuit that corrects the tap gain correction circuit, the correlator being a first Fourier transform circuit that performs the Fourier transform of the sample data stored in the first storage circuit. A second Fourier transform of the sample data stored in the second storage circuit
Fourier transform circuit, a conjugation circuit that receives the output of the first or second Fourier transform circuit and performs conjugation thereof, and to the conjugation circuit of the first or second Fourier transform circuit A multiplication circuit that is provided with the output of the Fourier transform circuit on the side that did not supply the output and the output of the conjugation circuit and performs multiplication between them, and an inverse Fourier transform circuit that performs the inverse Fourier transform of the output of the multiplier circuit And consist of
A ghost elimination device characterized by performing a correlation calculation on a frequency axis. 2. The ghost removing apparatus according to claim 1, wherein the correlator performs a Fourier transform on a cross-correlation operation between sample data taken out from the first and second memory circuits. 2. A ghost elimination device comprising: a microcomputer that is executed by software on the frequency axis in 3); and a third memory circuit that stores a program and output data of the microcomputer.